C＃中的HTML解析-分类

Question

I am working on a sentiment classification and I was parsing data from local movie database. Problem is that they have three forms of classification. One with stars (realized in ...) one "rubbish" and without giving stars or calling it rubbish here is the main link to it : http://www.csfd.cz/film/7049-playgirls/?all=1 you need to check source code - here is an example you can see all three kinds of user evaluation of movies.

</li>
<li id="comment-8356897">
    <h5 class="author"><a href="/uzivatel/138463-campbell/">Campbell</a></h5>
    <img src="http://img.csfd.cz/assets/images/rating/stars/2.gif" class="rating" width="16" alt="**" />
    <div class="info">
        <a href="/uzivatel/138463-campbell/komentare/">všechny komentáře uživatele</a></div>
    <p class="post">Ale jo:-D Když jsem viděl že tenhle film je na prvním místě mezi největšíma sračkama na CSFD, a tak jsem se zhrozil a abych si utrpení ještě vylepšil, tak jsem si pustil oba dva díly naráz. No hell to celkem bylo ale ne nic extrémní. Viděl jsem větší shity. V tomhle filmu jsem měl děsnej problém fandit někomu fandit protože to moc nejde. Šílenost, Ale ne nejhorší.<span class="date desc">(11.3.2011)</span></p>
</li>
<li id="comment-872277">
    <h5 class="author"><a href="/uzivatel/48974-fleker/">fleker</a></h5>

    <div class="info">
        <a href="/uzivatel/48974-fleker/komentare/">všechny komentáře uživatele</a></div>
    <p class="post">tak na todle rači ani koukat nebudu; hodnocení to má slušný ale nechci riskovat aby mi vyschla mícha<span class="date desc">(29.7.2009)</span></p>
</li>
<li id="comment-327360">
    <h5 class="author"><a href="/uzivatel/41698-ozo/">Ozo</a></h5>
    <strong class="rating">odpad!</strong>
    <div class="info">
        <a href="/uzivatel/41698-ozo/komentare/">všechny komentáře uživatele</a></div>
    <p class="post">Změna názoru - tohle si jednu hvězdičku nezaslouží =(<span class="date desc">(29.7.2007)</span></p>
</li>

Thanks a lot my plan was to do it like this :

string srxPathOfCategory = "//ul[@class='ui-posts-list']//li//img[@class='rating'] | //ul[@class='ui-posts-list']//li//strong[@class='rating']";
        foreach (var att in doc.DocumentNode.SelectNodes(srxPathOfCategory)) // | .//strong[@class='rating']")){
        {

            if (att.InnerText == "odpad!")  //odpad means rubbish
            {
                b[j] = att.InnerText; //saving "odpad!" for later use

            }
            if (att.Attributes["alt"] != null)

            {
                b[j] = att.Attributes["alt"].Value; //these values are from 1* to 5*****

            }
          if (att.InnerText != "odpad!" && att.Attributes["alt"] == null)//this is where the problems starts
            {
                   b[j] = "without user evaluation";

            }

            j++;
        }

Problem with this code is that if it fails to find att.InnerText == "odpad!" or att.Attributes["alt"] != null it continues to the next post and take user evaluation from there. But I would like to match at least something to the post where the evaluation was ommited.

Answer 1

thanks for all help but the problem was in tha xpath for html.

I solved it like this

string srxPathOfCategory = "//ul[@class='ui-posts-list']//li";

        foreach (var att in doc.DocumentNode.SelectNodes(srxPathOfCategory))
        {

            foreach (var child in att.ChildNodes.Skip(3)) // skipping first three nodes //- first one is whitespace - marked as #text child node, then there is h5 and third is //another whitespace marked as #text child node 
            {

                if (child.InnerText == "odpad!")
                {
                    b[j] = child.InnerText;
                    Console.WriteLine(b[j]);
                    Console.ReadKey();
                    break;

                }
                else if (child.Attributes["alt"] != null)
                {
                    b[j] = child.Attributes["alt"].Value;
                    Console.WriteLine(b[j]);
                    Console.ReadKey();
                    break;
                }
                else
                {
                    b[j] = "without user evaluation";
                    Console.WriteLine("hlupost");
                    Console.ReadKey();
                    break;
                }

            }
            j++;
        }

Answer 2

"odpad!" is not in an Attribute, it's in an Element.

Answer 3

What if you change your if statements. Why do you even have there 3 if statements if only one can be true?

// Is it "odpad" ?
if (att.InnerText == "odpad!")
{
    b[j] = att.InnerText;

}
// .. If not, is it starred?
else if (att.Attributes["alt"] != null)
{
    b[j] = att.Attributes["alt"].Value;

}
// If none of above, it must be this (default)
else
{
       b[j] = "without user evaluation";

}